Ive been putting together a presentation on how to build and run a PAT team, focusing on the areas that can go wrong, and the pitfalls to avoid. This process reminded me of several recent experiences with pharmaceutical company clients and an old R-rated joke no offense intended (irate readers, feel free to contact me).
Here goes (apologies in advance to little old ladies everywhere).
But lady, we dont have any lettuce, replied the grocer. She then proceeded to ask for other combinations of vegetables, always including lettuce. The exasperated grocer responded with a series of questions. How many Os are there in tomato? and How many Ts in potato? The lady answered each question correctly without any hesitation.
Finally, he asked, How many #$%! are there in lettuce? The lady answered, There is no #$%! in lettuce. Thats what Ive been telling you, the grocer replied. There is no #$%!in lettuce!
Well, there is an A, and a T in Process Analytical Technologies (PAT), but they are not the most important aspects of PAT. As Ajaz Hussain has always said, PAT is not about the hardware.
Yet, almost everywhere I visit, I meet people who think that sticking some analytical instrument nearby or into the process stream is running PAT. They buy a near-infrared or Raman spectrometer and declare, Im doing PAT!
Adding some analytical instrument to a process is no more PAT than adding an auto-sampler to an HPLC is automating the analysis. However, I know of some nutraceutical companies who get a stay of execution on their FDA inspections by simply stating that they are buying a NIR for raw material analysis. Therefore, is it any wonder that so many people think the act of buying a tool is sufficient?
The first part of PAT is, of course, the act of measurement, just as buying running shoes is part of running a marathon. You might need to wear shoes to run, but that is not, in itself, running. Similarly, you will need an instrument to take readings, but sticking a probe into some process machine is hardly PAT.
Why? Measurement merely gives you data. Whats wrong with that? Well, the fact that it is raining in Burma is data; as is the fact that chickens have feathers. Will either of these factoids influence your tablet production or analysis? Not unless you are making chicken-feather tablets in Burma.
Several years ago while consulting at a large company and working my NIR magic, an employee came into the lab and changed the graph paper in a temperature-humidity measuring device. I asked (out of curiosity) what he did with the data. He answered that he put the charts into a drawer. Of course, but what do you do with them? I asked.
Two full-time employees faithfully changed that paper every day and filed the graphs away. However, nothing was done with the reams of data generated. No temperature or humidity changes were ever made to the HVAC, and no analysis was performed that would have provided some information about the process.
The first step in PAT is taking data to understand the process under investigation. This data is used to correlate with the final product performance and understand its effects. For example, we can see what particle size, moisture or morphology of the raw materials does to blend uniformity and, ultimately, to such things as dissolution and in vivo performance. Monitoring a blend for uniformity is merely automating a system. Using the data to understand the effects of raw materials on blending is PAT.
Once the system is understood, it can be repeated on a regular basis. For example, the properties of a good blend must be seen before the mix is granulated. It follows that a good granulation must be seen before it is lubricated. Once we can establish that a particular step is understood, and under control, then we can reach for the brass ring. (How many of you younger readers remember brass rings on merry-go-rounds?) We may now make improvements to the product. This is the prize of PAT. As I stated, merely buying an instrument is no guarantee of an improved product, any more than buying $250 running shoes guarantees that youll make it through, much less win, the New York Marathon.
However, doing PAT correctly has an added bonus. Once you understand one product, you get a leg up on all the others made with a similar formulation. This always makes the first PAT program the most expensive to implement, with costs decreasing for each succeeding program.
Any questions? Therell be a quiz on this soon.